⚠️ La chute d'Oracle vient de déclencher le CRASH IA! (C'est terminé!)

Summary

The video argues that Oracle’s large AI-infrastructure bet could create financial risks beyond the company, particularly if OpenAI cannot meet its future commitments. It contrasts Oracle’s reported growth and record order book with rising capital spending, debt, and dependence on a major customer.

The potential market-wide consequences are presented as a hypothetical downside scenario, not a forecast.

Oracle’s AI strategy and financial position

Oracle is described as a long-established database company that fell behind Amazon Web Services and Microsoft Azure in cloud computing, then pivoted toward providing infrastructure for AI.

The video says Oracle signed a five-year, $300 billion cloud contract with OpenAI. Oracle’s total order book reportedly reached $664 billion, including $364 billion attributed to other customers.

After results announced in September 2025, Oracle shares reportedly rose from $240 to $345 in one session—about 40%—before subsequently falling. The video opens by claiming that the shares were down 60% in a year.

The latest quarterly results discussed showed:

• Revenue of $19.3 billion, up 30%.
• Cloud infrastructure revenue doubled year over year.
• Earnings per share rose 55%.

The video also cites a projected net margin above 25% in 2026, while emphasizing that near-term spending is running ahead of cash generation.

Other financial details cited include:

• Oracle’s cash reserves reportedly fell by $23.7 billion over the fiscal year.
• Capital expenditure was $55 billion in FY2026, with management expected to spend $70 billion in 2027.
• Debt rose from $92 billion to $129 billion in 12 months.
• Oracle announced plans to issue $20 billion in new shares, which could dilute existing shareholders.
• Oracle’s 30-year bonds were said to yield more than 8%, and its credit-default swaps (CDS) reached a record high. The video interprets this as investors demanding more compensation for Oracle’s credit risk.

Project Jupiter and financing risk

Oracle’s Stargate-related Project Jupiter, described as a roughly $165 billion data-center project in New Mexico, is central to the investment thesis.

The planned facility is described as requiring substantial power: 4.5 GW for the broader computing capacity, while a later passage cites 2.5 GW for the site. The video says Oracle plans to use natural-gas generation and needs a 17 km pipeline across public land.

A permit dispute, environmental concerns, and local opposition are said to be delaying the pipeline. The video reports that Oracle sent a force majeure notice to Blue Owl on September 24, 2026, signaling concern about construction delays.

Blue Owl is described as financing and building the site, then leasing it to Oracle for 18 years, at an estimated total cost of $35–55 billion. Once operational, rent would be about $3 billion a year, before Oracle’s own server investment.

The project’s financing reportedly includes $18 billion borrowed by around 20 banks. Following the delay notice, the debt was said to trade at 90% of face value.

The video also says Larry Ellison pledged more than $9 billion of Oracle shares as collateral for personal borrowing, presenting this as another potential warning sign.

OpenAI as a major customer

The video says more than a third of Oracle’s orders come from OpenAI, which is expected to pay Oracle approximately $60 billion annually from 2028 under the cloud agreement.

OpenAI is described as having generated about $12 billion in revenue in 2025, with $25 billion targeted for the current year and a higher possible figure of $70 billion. The video says it could lose around $38 billion in the current year and is spending about $1.50 for every $1 of revenue.

The video argues that OpenAI would need rapid growth and successful monetization of AI products to afford the Oracle payments.

Other companies with exposure to OpenAI or its AI infrastructure include:

• Nvidia, which reportedly announced up to $100 billion of investment.
• SoftBank, said to have committed around $40 billion.
• Microsoft, which the video says had invested more than $13 billion.

The video notes the opposing view that OpenAI may be “too big to fail” and mentions a possible Microsoft acquisition as one alternative being discussed. These are presented as investor opinions, not settled outcomes.

Hypothetical downside scenario

The video outlines the following chain as a possible scenario, explicitly saying it is not the most likely outcome:

1. OpenAI’s AI-agent and business monetization plans fall short, while competing models from China—particularly DeepSeek and Qwen—remain adequate and inexpensive.
2. OpenAI’s revenue grows by 40% annually rather than the expected 200%, making further fundraising difficult.
3. OpenAI seeks to reduce or reschedule its Oracle commitments—for example, lowering planned annual payments from $60 billion to $30 billion over a longer period.
4. Oracle’s order book falls by $200 billion, its shares drop 25%, and its credit rating is downgraded to speculative grade.
5. The video estimates that $30–40 billion of Oracle debt could be forced onto the market, with refinancing costs rising to 12–13%.
6. Further losses could affect lenders, private-credit investors, Nvidia, and broader equity markets. In this hypothetical, the S&P 500 falls, U.S. portfolios lose 15% in a month, and the CAC 40 declines 12%.

Portfolio exposure and recommendations

The video argues that U.S. index investors may have more exposure to AI-related technology stocks than they realize:

• Around 10 technology stocks are said to represent 40% of the S&P 500.
• An MSCI World ETF is said to contain more than 25% in the same securities.

It specifically warns that investors holding unit-linked life insurance, a PEA, a World ETF, or a retirement plan may be indirectly exposed to the AI trade.

The video suggests reviewing portfolios by:

• Checking the underlying holdings and concentration of funds.
• Considering rebalancing if 60–70% of a portfolio is unintentionally invested in U.S. technology stocks.
• Considering European or Asian financials and consumer-goods companies, excluding China, as well as gold and bonds.

It advises not to panic, while stressing that Oracle’s construction schedule and OpenAI’s ability to deliver on its business and funding plans remain uncertain.

Assets, companies, and instruments mentioned

• Equities and companies: Oracle, OpenAI, Nvidia, Amazon/AWS, Microsoft/Azure, SoftBank, Alphabet, Cisco, Intel, and Renault.
• AI competitors: DeepSeek and Qwen.
• Indices and funds: S&P 500, CAC 40, MSCI World, and World ETFs.
• Other instruments and assets: Oracle bonds, CDS, new share issuance, private credit, gold, bonds, U.S. and European equities, crypto including Bitcoin, and cash in euros and U.S. dollars.

No stock ticker symbols are explicitly given.

Sponsored-platform promotion and disclosures

The video includes a paid promotional segment for Neverless. It advertises access to U.S. shares, cryptocurrencies, gold, and silver; no trading fees; selected 24/7 trading; DCA, stop-limit orders, leverage up to 5×, and short positions.

The promotion also claims regulatory authorization and deposit-protection arrangements, and advertises daily cash interest of 1.84% on euros, 2.78% on U.S. dollars, and 2.42% or more on crypto. It offers €20 for a deposit of at least €100 held for 30 days. These are promotional claims made in the video and are not independently verified here.

No explicit “not financial advice” disclaimer appears in the supplied subtitles. The presenter does caution that the crash scenario is uncertain and is not considered the most likely outcome.

Presenter and sources

• Presenter: Eléonor, from MoneyRadar. The channel name is garbled in parts of the subtitles.
• Sources named or discussed: Oracle results and press releases; the reported force-majeure notice to Blue Owl; bond and CDS market pricing; and statements attributed to Oracle management, OpenAI, Nvidia, and other companies. The video also refers viewers to a source document linked in its description.